Automatic Quality Assessment of Transperineal Ultrasound Images of the Male Pelvic Region, Using Deep Learning.

Ultrasound guidance is not in widespread use in prostate cancer radiotherapy workflows. This can be partially attributed to the need for image interpretation by a trained operator during ultrasound image acquisition. In this work, a one-class regressor, based on DenseNet and Gaussian processes, was implemented to automatically assess the quality of transperineal ultrasound images of the male pelvic region. The implemented deep learning approach was tested on 300 transperineal ultrasound images and it achieved a scoring accuracy of 94%, a specificity of 95% and a sensitivity of 92% with respect to the majority vote of 3 experts, which was comparable with the results of these experts. This is the first step toward a fully automatic workflow, which could potentially remove the need for ultrasound image interpretation and make real-time volumetric organ tracking in the radiotherapy environment using ultrasound more appealing.

Centrifugal pump performance during low-flow extracorporeal CO2 removal; safety considerations.

AIM: The aim of this study was to examine the hydrodynamic performance and gaseous microemboli (GME) activity of two centrifugal pumps for possible use in low-flow extracorporeal CO2 removal. MATERIALS & METHODS: The performance of a Rotassist 2.8 and a Rotaflow 32 centrifugal pump (Maquet Cardiopulmonary AG, Hirrlingen, Germany) was evaluated in a water-glycerine mixture-filled in vitro circuit that enabled measurement of pressures and GME at the pump inlet and pump outlet. Pressure-flow curves were acquired in a 1,000 to 5,000 rpm range while increasing drainage resistance in one series and outlet resistance in another. RESULTS: Respective minimum pump inlet and maximum pump outlet pressures were -539 mmHg and 754 mmHg for the Rotassist 2.8 and -606 mmHg and 806 mmHg for the Rotaflow 32. Maximum standard deviations on pump pressures and flow amounted to 3.0 mmHg and 0.03 L/min, respectively, regardless of pump type and drainage or outlet resistance. The GME at the pump outlet were detectable at pump inlet pressures below -156 mmHg at 0.2 L/min and 2,500 rpm for the Rotassist 2.8 and below -224 mmHg at 0.9 L/min and 3,000 rpm for the Rotaflow 32. CONCLUSION: Both the Rotassist 2.8 and Rotaflow 32 centrifugal pumps show a comparably high hydrodynamic stability, but potential GME formation with decreasing pump inlet pressures should be taken into account to ensure safe centrifugal pump-based low-flow extracorporeal CO2 removal.

The visually-evoked cerebral blood flow response in women with a recent history of preeclampsia and/or eclampsia.

Several studies provide evidence for altered cerebral hemodynamics during (pre)eclampsia. Whether (pre)eclampsia has a persistent negative impact on cerebral hemodynamics, possibly contributing to an elevated risk of premature stroke, is unknown. The aims of this study were (i) to refine and apply a control system-based method previously introduced by Rosengarten to quantify the visually-evoked blood flow response of the posterior cerebral artery (PCA); and (ii) to test the hypothesis with this method that cerebral hemodynamics in women with a recent history of (pre)eclampsia is abnormal relative to that in parous controls. Hereto, we recorded cerebral blood flow velocity (CBFV) in the PCA by transcranial Doppler (TCD) sonography during cyclic visual stimulation in 15 former preeclamptics, 13 former eclamptics and 13 controls. The typical CBFV response was fitted with the step response of a second-order-linear model enabling quantification by parameters K (gain), zeta (damping), omega (natural frequency), T(v) (rate time) and T(d) (time delay). The method refinement introduced here consisted of response filtering before quantification and of considering the individual instead of group-averaged response patterns. Application of this refinement reduced the fitting errors (1.4 +/- 1.2 vs. 3.2 +/- 1.8, p < 0.01). Intergroup differences in model parameters were not found. Although statistically not significant, a trend was observed that critical damping (zeta>1) occurred more frequently in the combined group of former patients than in the controls (7 of 28 vs.1 of 13, p = 0.16). Critical damping (zeta>1) reflects an abnormal response, which is either compensated for by a rise in rate time ("intermediate"; zeta>1; T(v) > 20) or remains uncompensated ("sluggish"; zeta>1; T(v) < 20). Critical damping increased significantly (p = 0.039) with (pre-)eclampsia-to-test-interval in the PE+E patients with abnormal responses (zeta>1), suggesting that (pre)eclampsia might induce diminishing cerebral hemodynamic function over time. Based on a system-analytical classification approach, the data of this study provide evidence for individual CBFV responses to be abnormal in former (pre)eclamptics compared with controls. Further study is needed to reveal how the abnormal CBFV response classification reflects cerebrovascular dysfunction.